Update import paths

aquilign/align/bertalign/Bertalign.py

@@ -5,7 +5,7 @@
 __author__ = "Jason ([email protected])"
 __version__ = "1.1.0"
 
-from bertalign.encoder import Encoder
+from aquilign.align.bertalign.encoder import Encoder
 # from sonar.inference_pipelines.text import TextToEmbeddingModelPipeline
 # See other cross-lingual embedding models at
 # https://www.sbert.net/docs/pretrained_models.html
@@ -19,4 +19,4 @@
 # nb = input(f'Choose a model:')
 model = Encoder(models[int(1)])
 
-from bertalign.aligner import Bertalign
+from aquilign.align.bertalign.aligner import Bertalign

aquilign/align/bertalign/aligner.py

@@ -1,8 +1,8 @@
 import numpy as np
 
-from bertalign.Bertalign import model
-import bertalign.corelib as core
-import bertalign.utils as utils
+from aquilign.align.bertalign.Bertalign import model
+import aquilign.align.bertalign.corelib as core
+import aquilign.align.bertalign.utils as utils
 import torch.nn as nn
 import torch
 

aquilign/align/bertalign/encoder.py

@@ -1,7 +1,7 @@
 import numpy as np
 import torch
 from sentence_transformers import SentenceTransformer
-from bertalign.utils import yield_overlaps
+from aquilign.align.bertalign.utils import yield_overlaps
 # from sonar.inference_pipelines.text import TextToEmbeddingModelPipeline
 
 

aquilign/align/bertalign/syntactic_tokenization.py

@@ -5,7 +5,7 @@
 import json
 import sys
 import langid 
-import bertalign.utils as utils
+import aquilign.align.bertalign.utils as utils
 
 def syntactic_tokenization(path, corpus_limit=None, use_punctuation=True):
     name = path.split("/")[-1].split(".")[0]
@@ -16,7 +16,7 @@ def syntactic_tokenization(path, corpus_limit=None, use_punctuation=True):
     codelang, _ = langid.classify(text[:300])
     print(text)
     print(codelang)
-    with open("bertalign/delimiters.json", "r") as input_json:
+    with open("aquilign/align/bertalign/delimiters.json", "r") as input_json:
         dictionary = json.load(input_json)
     # Il ne reconnaît pas toujours le castillan
     if codelang == "an":

main.py

@@ -6,10 +6,10 @@
 import sys
 import numpy as np
 # import collatex
-import graph_merge
-import bertalign.utils as utils
-import bertalign.syntactic_tokenization as syntactic_tokenization
-from bertalign.Bertalign import Bertalign
+import aquilign.align.graph_merge as graph_merge
+import aquilign.align.bertalign.utils as utils
+import aquilign.align.bertalign.syntactic_tokenization as syntactic_tokenization
+from aquilign.align.bertalign.Bertalign import Bertalign
 import pandas as pd
 import argparse
 import glob
@@ -81,6 +81,10 @@ def __init__(self, corpus_size:None,
         self.use_punctiation = use_punctuation
         self.prefix = prefix
 
+        try:
+            os.mkdir(f"result_dir")
+        except FileExistsError:
+            pass
         try:
             os.mkdir(f"result_dir/{self.out_dir}/")
         except FileExistsError:
@@ -219,7 +223,7 @@ def run_alignments():
     prefix = args.prefix
     use_punctuation = args.use_punctuation
     print(f"Punctuation for tokenization: {use_punctuation}")
-    MyAligner = Aligner(corpus_size=None, max_align=3, out_dir=out_dir, use_punctuation=use_punctuation, input_dir=input_dir, main_wit=main_wit, prefix=prefix)
+    MyAligner = Aligner(corpus_size=100, max_align=3, out_dir=out_dir, use_punctuation=use_punctuation, input_dir=input_dir, main_wit=main_wit, prefix=prefix)
     MyAligner.parallel_align()
     utils.write_json(f"result_dir/{out_dir}/alignment_dict.json", MyAligner.alignment_dict)
     align_dict = utils.read_json(f"result_dir/{out_dir}/alignment_dict.json")